Estimation of breast, prostate, and colorectal cancer incidence using a French administrative database (general sample of health insurance beneficiaries)

Aim. - The aim of this study was to compare incidence of breast, prostate, and colorectal cancer incidence estimated from a French administrative database with the incidences estimated from the cancer registry data. Materials and methods. - A cohort of 426,410 people included in the general sample of health insurance beneficiaries (EGB) database as of January 1, 2007, was constituted. Several algorithms were developed to estimate cancer incidence between 2008 and 2012 using principal diagnosis (PD) of hospital discharge data (medical information systems program [PMSI]) and/or long-term disease (LTD) and together with a procedure necessary for histological diagnosis and indicating initial disease management. The incidence rates obtained were compared with those from the registry data using the standardized incidence ratio (SIR). Results. - The algorithm taking into account LTD and PD in the PMSI and the mandatory presence of a marker procedure provided estimates close to those from the registry data for breast cancer (SIR: 1.12 [1.07-1.18]) and colorectal cancer (SIR: 0.94 [0.88-1.02] in men and SIR: 0.93 [0.86-1.01] in women). For prostate cancer, taking into account specific procedures and drugs in addition to LTD and PD in the PMSI enhanced the estimation of incidence (SIR: 1.03 [0.98-1.08]). Conclusion. - The PMSI together with reimbursement data (LTD, procedures, drugs) provided estimates of breast, prostate, and colorectal cancer incidence, at a national level, comparable to those from the cancer registry data. (C) 2016 Elsevier Masson SAS. All rights reserved.